check_procs -w 10 -c 20 --metric=CPU parameter invalidation


I test ‘check_procs’ of - metric parameters, the use of the CPU and PROCS(default) effects, who can help solve the problem?



–metric CPU checks and alerts agaisnt any process that has it’s current percentage CPU utilisation above the warning/critical thresholds

[root@localhost libexec]# ./check_procs -m CPU -w 40 -c 50 
CPU OK: 84 processes
[root@localhost libexec]# ./check_procs -m CPU -w 4 -c 50 
CPU WARNING: 4 warn out of 110 processes
[root@localhost libexec]# ./check_procs -m CPU -w 4 -c 8
CPU CRITICAL: 4 crit, 0 warn out of 103 processes

–metric PROCS (or indeed no specified metric) will alert against the current total number of processes running compared to the thresholds.

[root@localhost libexec]# ./check_procs -m PROCS -w 200 -c 400 
PROCS OK: 99 processes
[root@localhost libexec]# ./check_procs -m PROCS -w 10 -c 400 
PROCS WARNING: 114 processes
[root@localhost libexec]# ./check_procs -m PROCS -w 10 -c 20 
PROCS CRITICAL: 184 processes

Of course, if you don’t specify a threshold, you’ll always get an OK status returned in either case.




When -m for the CPU, ok state, showed that the same process, but to occupy CPU - w value/ warning, - c value/ alerts.